@axolotl I agree some interference with unethical endeavours might make sense! I just don't like the blanket dismissals because it means we don't look at it on a per-case basis, nor consider the tradeoffs.
For example: I think it's fine-ish to use models from Mistral (French AI), or use some of the already reasonable local models -- in particular if what we *do* with them is prosocial (e.g. create tools that democratize information access for the people, or enable more people to improve their lives in some quantifiable way). And from there we can argue where we draw the line?
